Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/231350 
Authors: 
Year of Publication: 
2021
Series/Report no.: 
DoCMA Working Paper No. 5
Publisher: 
TU Dortmund University, Dortmund Center for Data-based Media Analysis (DoCMA), Dortmund
Abstract: 
The current coronavirus pandemic is a recent example of how an unpredictable event can cause uncertainty and affect people, markets and economies worldwide. Economic indicators like the well-known and frequently cited Economic Policy Uncertainty Index, developed by the three economists Scott R. Baker, Nicholas Bloom and Steven J. Davis, try to capture perceptions of uncertainty and use them to generate predictions for the economic future. The index is based on a search term that is employed to select relevant elements from a population of articles. However, previous research contributions criticize the composition of the search term for being too broad and restricted to specific policy areas, thereby missing detections of new or unforeseen sources of economic uncertainty. This research note aims to modify the economists' search term in three different ways. The author introduces the term and concept of "risk", which is not considered in the original search term, adds further policy areas that can cause economic uncertainty, and extends the term "uncertainty" by including related terms and synonyms. In order to evaluate the success of the optimized search terms in selecting relevant articles from a population of articles, all search terms are applied to two randomly drawn samples, derived from an original corpus of 2,723,049 articles and a pre-filtered corpus of 514,297 articles from the German daily newspapers Handelsblatt and Süddeutsche Zeitung. The investigation period ranges from January 1994 to March 2020. For comparison, both samples are also filtered using Baker, Bloom and Davis's original search term. The different selection results are evaluated with the help of the parameters recall and precision. The results are preliminary, but encouraging. In both samples, only around every tenth relevant article is selected when Baker, Bloom and Davis's search term is applied. By modifying the original search term, the recall could be increased considerably with little disadvantages in terms of precision. The research process shows that economic uncertainties can be related to other concepts and policy areas that are not captured by the economists' original search term.
Subjects: 
economic policy
uncertainty
risk
search term
optimization
media analyses
recall
precision
Persistent Identifier of the first edition: 
Document Type: 
Working Paper

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.